daikin-industries-ltd/ja-fineweb-2-hvac-fastText-filtered-v3

该数据集是一个经过fastText分类器筛选的日语文本数据集，专门针对HVAC（空調・暖房・換気・空気調和）相关的内容。数据集来源于llm-jp/llm-jp-corpus-v3的ja_fineweb_2数据，通过fastText分类器筛选出标签为__label__positive的记录。数据集包含790,289条记录，总字符数为970,505,001，平均每条记录1,228.0个字符。每条记录包含text、meta、fasttext_label和fasttext_score四个字段。

This dataset is a filtered Japanese text dataset using a fastText classifier, specifically targeting HVAC (heating, ventilation, air conditioning) related content. The dataset is derived from the ja_fineweb_2 data of llm-jp/llm-jp-corpus-v3, filtered to include only records labeled as __label__positive by the fastText classifier. The dataset contains 790,289 records with a total of 970,505,001 characters, averaging 1,228.0 characters per record. Each record includes four fields: text, meta, fasttext_label, and fasttext_score.